Visible to the public The Shadow Nemesis: Inference Attacks on Efficiently Deployable, Efficiently Searchable Encryption

TitleThe Shadow Nemesis: Inference Attacks on Efficiently Deployable, Efficiently Searchable Encryption
Publication TypeConference Paper
Year of Publication2016
AuthorsPouliot, David, Wright, Charles V.
Conference NameProceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security
Conference LocationNew York, NY, USA
ISBN Number978-1-4503-4139-4
KeywordsCollaboration, composability, efficiently deployable efficiently searchable encryption, encrypted email, Encryption, encryption audits, Human Behavior, ios, iOS encryption, Metrics, pubcrawl, Resiliency, Scalability, Searchable encryption, security

Encrypting Internet communications has been the subject of renewed focus in recent years. In order to add end-to-end encryption to legacy applications without losing the convenience of full-text search, ShadowCrypt and Mimesis Aegis use a new cryptographic technique called "efficiently deployable efficiently searchable encryption" (EDESE) that allows a standard full-text search system to perform searches on encrypted data. Compared to other recent techniques for searching on encrypted data, EDESE schemes leak a great deal of statistical information about the encrypted messages and the keywords they contain. Until now, the practical impact of this leakage has been difficult to quantify. In this paper, we show that the adversary's task of matching plaintext keywords to the opaque cryptographic identifiers used in EDESE can be reduced to the well-known combinatorial optimization problem of weighted graph matching (WGM). Using real email and chat data, we show how off-the-shelf WGM solvers can be used to accurately and efficiently recover hundreds of the most common plaintext keywords from a set of EDESE-encrypted messages. We show how to recover the tags from Bloom filters so that the WGM solver can be used with the set of encrypted messages that utilizes a Bloom filter to encode its search tags. We also show that the attack can be mitigated by carefully configuring Bloom filter parameters.

Citation Keypouliot_shadow_2016